
OpenAI Launches Health in ChatGPT - Connect Medical Records and Apple Health Across Every Chat
300 million people a week already ask ChatGPT health questions. Now those answers can pull from your actual lab results, medications, and sleep data - not just population-level knowledge.
OpenAI launched Health in ChatGPT today - a feature that lets logged-in U.S. users on Free, Go, Plus, and Pro plans connect Apple Health and medical records from supported hospital systems, One Medical, and Function Health. Available now on web and iOS for users 18 and older. Health in ChatGPT lets connected data appear across all regular conversations, not just inside a dedicated sidebar section. Health is not available in Codex.
70% of Health Questions Came From Outside the Health Section
OpenAI previously tested Health with a limited group that had to visit a separate section of ChatGPT to get the full benefit. Early data showed more than 70% of health-related conversations took place outside that dedicated space. That result drove the redesign. Connected health context now follows users across all conversations without requiring a detour to a sidebar - ChatGPT can consider a dietary restriction when suggesting a restaurant, or factor in a recent injury when planning the weekend. Add @Health to any message to explicitly pull in health data, or let ChatGPT ask for permission each time.
What You Can Connect
Apple Health brings sleep, activity, workouts, and nutrition data from apps that share into Apple's framework - though proprietary scores from specific fitness apps may not transfer. Medical records from U.S. hospital systems connect over FHIR, the standard protocol for health data exchange. Apple rebuilt Siri with Google Gemini at WWDC 2026, putting Apple Health data at the center of two AI assistant platforms now. ChatGPT can compare lab results across visits, summarize what changed since a last appointment, and flag patterns worth raising with a doctor. Records may lag reality: a medication listed after you stop taking it won't automatically disappear, and OpenAI recommends checking important data against the original source.
Privacy: Not Used for Training - But Check the Deletion Terms
Connected medical records and Apple Health data are not used to train OpenAI's foundation models or target ads, regardless of a user's standard model-training setting. Conversations that draw on health data get additional encryption beyond standard ChatGPT protections. OpenAI does not create memories directly from medical records or Apple Health data, though health conversations can generate them. A chatbot reading your actual lab results is more useful than one answering from general population knowledge - but that usefulness depends entirely on whether users trust OpenAI to store their medical history responsibly.
Disconnecting a data source removes synced records from OpenAI's systems within 30 days. Conversation history that already includes health information stays until a user manually deletes those chats. Before ChatGPT shares health context through another connected plugin - like sending a training plan based on Apple Health data to a running partner - additional safeguards check that the action matches what the user asked for.
GPT-5.6 Sol Handles Complex Health Reasoning on Paid Plans
GPT-5.6 Sol runs health conversations for paid users, while GPT-5.5 Instant handles health queries on the Free plan. OpenAI says every GPT-5.6 model outperformed GPT-5.5 on HealthBench Professional, an evaluation developed with hundreds of physicians worldwide. More than 300 million people a week already ask ChatGPT health questions - Health gives the model actual data to ground those answers instead of treating every question as if asked by someone with no prior medical context. OpenAI has not announced an EU, UK, or non-US launch date.